AVB-a, and describe two cones, of which V is the common vertex, passing through the two given conics. The sections of these cones made by any plane parallel to the plane VAB will be hyperbolas, of which the asymptotes are parallel to VA, VB respectively, and will therefore be similar and similarly situated to one another, and of the required form.

19. We now come to the most important and most difficult point of the theory of projections, the process by which, from the properties of the circle, those of conic sections in general may be deduced. We have just seen that any two conics may be projected into hyperbolas of any assigned eccentricity. Now this process, the possibility of which we have shewn by a geometrical method, of course admits of algebraical proof. And the algebraical investigation, on account of the continuity of the symbols employed, would not take any account of the restrictions introduced into the geometrical investigation, either as to the conics intersecting in real points, or as to the eccentricity of the conics into which they are projected being greater than unity. It is therefore possible, by an algebraical process, to transform the equations of any two conics whatever into those of conics of any eccentricity, and therefore into those of circles. The points and tangents common to the two given conics will be transformed into points and tangents common to their projections, and the relations of poles and polars will remain unaltered.

Since all circles pass through the same two points on the line at infinity, it follows that all circles are transformed by projection into a system of conics passing through the same two points, or having a common chord. Again, since every parabola touches the line at infinity, it follows that all parabolas will project into a system of conics touching the same straight line. A system of parabolas and circles will project into a system in which all the circles will become conics passing through the same two points, and all the parabolas will become conics, having the straight line joining those two points for a common tangent.

20. We have seen, in the investigation of the co-ordinates of the real and imaginary foci, given-in Chap. VI., that the pair

of imaginary tangents, drawn to a conic from any one of its four foci, satisfy the analytical conditions of being asymptotes to a circle. Hence these tangents must themselves meet the line at infinity in the two circular points. Conversely, if from the two circular points at infinity two pairs of tangents be drawn to any conic, these will form an imaginary quadrilateral, circumscribing the conic, the four angular points of which are the four foci of the curve.

Hence all conics having the same focus project into conics having a pair of common tangents; and all confocal conics into conics inscribed in the same quadrilateral.

The directrix is the polar of the focus, hence, if two conics have the same focus and directrix, they project into two conics having a common chord of contact for their common tangents, that is, having double contact with one another.

21. The anharmonic ratio of any pencil or range is unaltered by projection.

Let the transversal PQRS cut the four straight lines OP, OQ, OR, OS. Take any point V, not lying in the plane through these straight lines, join VO, VP, VQ, VR, VS, and let these lines be cut by any other plane in O', P', Q', R', S'.

Then

Hence the anharmonic ratio of the given pencil and range is the same as that of their projection.

22. The following proposition is useful in the projection of theorems relating to the magnitude of angles.

Any two lines which make an angle A with each other, form with the lines joining the circular points at infinity to their point of intersection, a pencil of which the anharmonic ratio is ε(24)√=1

It will be understood that the two given lines are taken as the first and third legs of the pencil.

Take the two lines as two sides of the triangle of reference, and let them be denoted by B=0, y=0. The lines joining their point of intersection to the circular points at infinity are given by eliminating a between the equation of the line. at infinity and that of the circumscribing circle, that is, between

form with B=0 and y = 0 a pencil of which the anharmonic

(Art. 23, Chap. 1.). In the present case,

COR. In the case in which the lines are at right angles

π

to one another, A =, and the anharmonic ratio becomes

2

unity, that is, the four lines form an harmonic pencil,

23. The known property of a circle, that "the angles in the same segment are equal to one another," gives rise to an important anharmonic property of conic sections. The property of the circle may be expressed thus, that "if A, B be any two fixed points on the circumference of a circle, O any moving point on it, the angle AOB is constant." Project the circle into any conic, and let A', B', O' be the projections of A, B, 0; H, K those of the circular points at infinity. Then, from the result of the last article, it follows that

{0'. A'B'HK) is constant.

Or, the anharmonic ratio of the pencil, formed by joining any point of a conic to four fixed points on the curve, is con

stant.

Reciprocating this theorem, in accordance with Art. 13, Chap. VII., we see that if any tangent to a conic be cut by four fixed tangents, the anharmonic ratio of the range, formed by the points of section, is constant.

24. If P, Q, R be three points in a straight line, and p, q, r be their projections, and s the projection of the point at infinity on the line PQR, then

where S denotes the point at infinity on the line PQR.

Also RS: PS in a ratio of equality, hence

PQ [pqrs] = QR'

25. If P, P, Q, Q, R, R... be a system of points in involution, and p, p, q, q, r, r... their projections, then since by Art. 27, Chap. 1. [PQRS] = [P'Q'R'S'], and by Art. 21 of this Chapter [PQRS] = [pqrs], [P'Q'R'S']=[p'q'r's'];;

it follows that [pqrs] = [p'q'r's'], or p, p, q, q, r, r'... are a system of points in involution. Hence, any system of points in involution projects into a system in involution.

If P coincides with P', p will coincide with p', or the foci of one system project into the foci of the other. We may observe that the centre of one system will not, in general, project into the centre of the other.

26. Let a system of circles be described through two given points A, A', and let any circle of the system cut a given straight line in P, P'. Produce AA' to meet the given straight line in O. Then

or OP. OP' is constant for all circles passing through A, A'. Hence, the system of points in which a system of circles, passing through two given points, cut a given straight line, are in involution. Project the system of circles into a system of conics, passing through four given points, and we learn that "a system of conics, passing through four given points, cut any straight line in a system of points in involution."

Of this system of conics, one can be drawn so that one of its points of intersection with the given straight line shall be at an infinite distance, in other words, so that one of its asymptotes shall be parallel to the given straight line. The other point, in which this conic cuts the given straight line, will be the centre of the system.

Again (see Art. 3, Chap. IX., infra), two conics can be described, passing through the four given points, and touching the given straight line. The two points of contact of these conics will be the foci of the system of points in involution.

By reciprocating these propositions, we obtain analogous properties of the system of conics, inscribed in a given quadrilateral, whence, by projection, may be obtained those of a system of confocal conics.

27. When the vertex of the cone, used for purposes of projection, is infinitely distant, so that the cone itself be